- Pronouns
- he / him / his
"Does It Matter When Presidential Candidates Post Their Platform?"HEM Tiberius
ENN Publisher
A longstanding "trick" I share with people is that Presidential candidates who post their platform later in the standing period have an inherent edge up over candidates who post their platforms earlier.
There's a logical thread to this theory. Basically, the public — excited for an election — will thoroughly read the first platform and direct all their election energy to critiquing it. By the time subsequent candidates post their manifestos, a fatigue has beset the electorate and they spend less time tearing into subsequent platforms. Result? The ticket who posted first gets dinged, and the ticket who posted second gets off much easier.
I decided to see if this old postulate stood up to the data. I went back through every competitive Presidential contest (competitive being defined as a race with two or more candidates on the ballot) through 2014 and recorded the following:
— Date standing was called
— Date each candidate posted their platform
— Final result of the election
I then used this data to make a scatterplot of the "how many days after standing was the platform posted" and how the candidate ultimately fared in the election.
The results of that scatterplot are not favorable to my old theory:
The Pearson correlation coefficient between how late into the election cycle a platform was posted, and the final result was 0.0097. To say "statistically insignificant" would be a massive understatement.
I decided to remix the data another way just to be sure. I took the winner of each Presidential contest, and pulled the number of days after his opponent that they posted their platform. If the winner posted their platform before the loser, that number would be a negative.
Both the average and median of this data were "0".
When plotting the number of days after opponent on a histogram, the data ever-so-slightly skewed right, but resembled a normal distribution.
When I removed the outliers (defined as [Q1 - IQR *1.5] and [Q3 + IQR * 1.5]) the histogram took an even greater resemblance to the normal distribution.
So, unfortunately for one of my pet theories, it appears as if the timing of posting a platform has no observable impact on election outcomes.
Of course, this mini-study cannot observe the "change in support" based on platform timing. It is possible that platform timing impacts support a great deal, but without pre-platform polling to use a frame of reference, it's impossible to tell.
As for now, it would seem that candidates have more things to worry about than when they post their platform.
##